Towards Enhanced Local Explainability of Random Forests: a Proximity-Based Approach